Introduction

According to the World Health Organization’s Global Status Report on Road Safety 2023, approximately 1.19 million people die each year as a result of road traffic crashes, making road traffic incidents the twelveth leading cause of death globally for all age groups1. The Global Burden of Disease reported an estimated 1.20 million deaths due to road injuries in 2021, with 3.2 times as many male deaths as female deaths2.

Glaucoma is characterized by progressive optic neuropathy, leading to irreversible visual field impairment (VFI)3 of varying severity. Globally, it remains a major cause of irreversible blindness, with an estimated 3.61 million individuals blind and 4.14 million experiencing visual impairment due to glaucoma as of 20204. Most glaucoma patients are asymptomatic5,6 and remain untreated7,8. Previous studies have reported that VFI due to glaucoma may increase the risk of motor vehicle accidents (MVAs)9,10,11,12,13,14,15. In a previous report, glaucoma patients had a 1.65 times higher rate of at-fault motor vehicle collisions in the past five years compared to those without glaucoma (95% CI, 1.20–2.28; P < 0.01)15. In addition, patients with visual impairment had higher incidence of collisions in driving simulator16. Therefore, early detection of glaucoma is crucial to prevent its progression and reduce the risk of MVAs caused by VFI.

A visual field (VF) self-check test may be a useful strategy for detecting asymptomatic glaucoma and helping people to recognize their own VFI, even in the early stages of the disease. The CLOCK CHART was developed as a self-check screening test for detecting VFI including glaucoma17. In addition, a previous study found that the near miss incidents related to VFI while driving may be risk factors for motor vehicle collisions among Japanese drivers18. The incident of VF-related near misses while driving may serve as a self-check test for detecting glaucoma in drivers. However, no studies have yet examined the association of VFI assessed using the CLOCK CHART (see Supplementary Fig. S1 online) and VF-related near miss incidents with MVAs among commercial drivers in a real-world setting. Although reasonable validation of the CLOCK CHART has been reported for patients with glaucoma, its usefulness for early detection of glaucoma remains unclear17.

We conducted two studies to address these gaps. In Study 1, we examined the association of the CLOCK CHART detected defects (CCDD) and VF-related near misses with MVAs among Japanese taxi drivers. In Study 2, we conducted a follow-up survey to assess the usefulness of the CLOCK CHART for detecting glaucoma among participants who had CCDD in Study 1.

Material and methods

Study 1

Participants

This cross-sectional study involved 1,921 workers, aged 20–74 years, employed by a Japanese taxi company in the Tokyo Metropolitan Area between August 2018 and March 2019. Workers who were not taxi drivers (n = 331), people with missing data (n = 255), and people with outlier variables such as driving distance (n = 2) were excluded. We also excluded women (n = 106) due to the small numbers of them, resulting in a final sample of 1,227 participants (Fig. 1). Trained staff explained the study protocol to the participants and each one provided their written informed consent. The study was approved by the Ethical Review Board of Juntendo University Faculty of Medicine (No. 2019059). All methods in this study were performed in accordance with the relevant guidelines and regulations.

Fig. 1
figure 1

Enrollment flowchart for Study 1.

Full size image

Measurement

Motor vehicle accidents

Data on MVAs were collected using a self-administered questionnaire. An MVA was defined as any accident in the past five years, including the following six categories: 1) single-car accidents with property damage caused by the taxi driver, 2) accidents with other vehicles without the fault status of the taxi driver, 3) accidents with other vehicles with the fault status of the taxi driver, 4) accidents with humans without the fault status of the taxi driver, 5) accidents with humans with the fault status of the taxi driver and the person hit, and 6) accidents with humans with the fault status of the taxi driver.

CLOCK CHART

The CLOCK CHART is a simple, paper-based visual field screening tool developed by Professor Chota Matsumoto of Kindai University17. It consists of a circular sheet of paper (40 cm in diameter) with four illustrated targets—a ladybug, a caterpillar, a butterfly, and a cat—placed at eccentricities of 10°, 15°, 20°, and 25°, respectively, in four different quadrants.

In the present study, trained staff members conducted the test following the protocol described by Matsumoto et al.17. The participant sat at a distance of 35 cm from the chart placed on a table. With the left eye covered, the participant was instructed to fixate on the red central dot with the right eye. The examiner first confirmed that the green caterpillar—located in the physiological blind spot (Mariotte blind spot)—was not visible. Then, the chart was slowly rotated clockwise by the examiner, with the participant maintaining central fixation. The rotation was paused every 30°, and the examiner recorded whether each target was visible. Fixation was also checked every 90°. The targets were evaluated in the following order of eccentricity: ladybug (10°), caterpillar (15°), butterfly (20°), and cat (25°). At the start of each eccentricity check, the blind spot was reconfirmed. The same procedure was repeated for the left eye. If any of the targets (excluding the blind spot) was reported as invisible in either eye, the participant was classified as having a CCDD. In accordance with the original validation study17, test points along the horizontal meridian were excluded from this analysis.

Visual field-related near miss

VF-related near miss incidents were assessed using a self-administered questionnaire. Based on a previous study18, the presence of a VF-related near miss was defined by one or more near-miss incidents in the previous five years, as indicated by the following nine statements: 1) I lost sight of traffic light that should have been there, 2) I lost sight of stop sign that should have been marked there, 3) A vehicle appeared suddenly from nowhere in front of me or disappeared, 4) A pedestrian appeared suddenly from nowhere in front of me or disappeared, 5) I found myself driving faster or slower than the normal traffic flow, 6) I find myself driving outside the lane, although I intend to stay within the lane, 7) I had difficulties in recognizing traffic signs, 8) I lost sight of where on the roadway I have been driving and get confused, 9) I have been pointed out by a passenger in car, such as a family member, that I was driving dangerously.

Covariates

A self-administered questionnaire was used to collect data on age (years), height (cm), body weight (kg), drinking status (current drinker or nondrinker), smoking status (current smoker or nonsmoker), daily working hours, hypertension diagnosis (yes or no), diabetes mellitus diagnosis (yes or no), previous year’s driving distance (km), and taxi driving experience. Excessive daytime sleepiness was defined as an Epworth Sleepiness Scale score of ≥ 1119. Body mass index (BMI) was calculated as the participant’s weight divided by the square of their height in meters.

Statistical analysis

We used Student’s t-tests or Mann–Whitney U tests to compare the participants’ demographics based on the presence of CCDD or VF-related near misses, along with χ2 tests for dichotomous variables. We used a multivariable-adjusted generalized linear model to examine the association of the presence of CCDD and VF-related near misses with MVA. We also examined the association of combination of both CCDD and VF-related near misses with MVA. Age, BMI, drinking status, smoking status, excessive daytime sleepiness, hypertension, diabetes mellitus, driving experience, annual driving distance, and daily working hours were used as confounding factors in these models. All p-values for statistical tests were two-tailed, with values < 0.05 considered statistically significant. SAS version 9.4 (SAS Institute, Cary, NC, USA) was used for all analyses.

Study 2

Participants

This study involved the 1,921 workers who participated in Study 1 and was conducted from July 2019 to February 2020. From this group, we extracted those who had CCDD yielding 326 participants who were included in the analysis (Fig. 2). The study was approved by the Ethical Review Board of Juntendo University Faculty of Medicine (No. 2019059). All methods in this study were performed in accordance with the relevant guidelines and regulations.

Fig. 2
figure 2

Participant flowchart for Study 2.

Full size image

Assessment

After conducting Study 1, we recommended ophthalmological checkups for participants with CCDD and sent them a self-administered questionnaire to determine whether they had received a diagnosis related to VFI, such as glaucoma, retinitis pigmentosa, or macular degeneration. The questionnaire included the following items: 1) whether or not they had received ophthalmological checkups (yes or no); 2) the date of the checkup (year, month, and day); and 3) the diagnosis (multiple choice: none, glaucoma, retinitis pigmentosa, macular degeneration, and other).

Results

Study 1

The proportions of participants with CCDD, VF-related near misses, and MVAs were 15.9%, 38.8%, and 61.9%, respectively (Table 1). The CCDD group had a higher mean age, more driving experience, and higher proportions of hypertension, diabetes mellitus, and VF-related near misses than the non-CCDD group (p < 0.05) (Table 2). The CCDD group also had a lower proportion of MVAs and current drinkers (p < 0.05). The group with VF-related near misses had a lower mean BMI and a higher proportion of MVAs and excessive daytime sleepiness than the group without VF-related near misses (p < 0.05, p < 0.01 and p < 0.01, respectively) (Table 3).

Table 1 Demographic characteristics of 1,227 Japanese male taxi drivers.
Full size table
Table 2 Demographic characteristics of 1,227 Japanese male taxi drivers based on the presence of the CLOCK CHART detected defects.
Full size table
Table 3 Demographic characteristics of 1,227 Japanese male taxi drivers based on their experience of visual field-related near misses.
Full size table

The multivariable-adjusted prevalence ratio (PR) (95% confidence interval [CI]) of MVAs for the CCDD group was 1.01 (0.90–1.14) compared to the non-CCDD group (Table 4). The multivariable-adjusted PR (95% CI) of MVAs for the group with VF-related near misses was 1.17 (1.08–1.28) compared to the group without VF-related near misses (Table 5). The multivariable-adjusted PRs (95% CI) of MVAs for the groups with only CCDD, only VF-related near misses, and both CCDD and VF-related near misses were 0.95 (0.79–1.14), 1.16 (1.06–1.27), and 1.22 (1.05–1.41), respectively, compared to the group with neither CCDD nor VF-related near misses (Table 6).

Table 4 Multivariable-adjusted prevalence ratios and 95% confidence interval for motor vehicle accidents among 1,227 Japanese male taxi drivers according to the presence of the CLOCK CHART detected defects.
Full size table
Table 5 Multivariable-adjusted prevalence ratios and 95% confidence interval for motor vehicle accidents among 1,227 Japanese male taxi drivers based on their experience of visual field-related near misses.
Full size table
Table 6 Multivariable-adjusted prevalence ratios and 95% confidence interval for motor vehicle accidents among 1,227 Japanese male taxi drivers based on the combination of the CLOCK CHART detected defects and visual field-related near misses.
Full size table

Study 2

Questionnaires were collected from 216 of the 326 participants (66.3%). Ninety-one of the 326 participants (27.9%) did not respond to the questionnaire, and 19 (5.8%) withdrew due to retirement during the study period. Among the 216 respondents, 71 (32.9%) visited an ophthalmologist, and 14 participants were diagnosed with glaucoma (Fig. 2).

Discussion

In Study 1, we identified a significant association between VF-related near miss incidents while driving and MVAs. We also found that the combination of VF-related near miss incidents and CCDD was associated with MVAs among male taxi drivers, even after adjusting for potential confounders.

In this study, 61.9% of the participants reported having MVA experience in the past five years. Due to varying definitions of MVA across studies20,21,22,23, making direct comparisons are difficult. Previous studies on taxi drivers reported the proportion of participants who had experienced MVAs as 9–56%21,22 over the previous two years and 23–26%20,23 over the previous three years. These studies used a shorter period for evaluating MVA history than this study, which used a period of five years, regardless of the severity of the incident. This longer assessment period presumably contributed to the higher reported proportion of MVA.

We did not observe a significant association between CCDD and MVAs (Table 4). Furthermore, in the combination analysis, participants with CCDD only did not have a significantly higher proportion of MVAs than those without CCDD or VF-related near misses (Table 6). The reason is unclear, but the risk of MVAs has been shown to increase with the severity of VFI, whereas mild VFI does not appear to contribute to MVAs13,15,24. Although, in this study, the severity of VFI was not assessed, this study was conducted in an occupational health setting, the proportion of patients with severe glaucoma may have been low. This may mask the association between VFI assessed by CLOCK CHART and MVAs in this study.

The ability of drivers to compensate for VFI through eye, head, or shoulder movements may mitigate their risk of traffic accidents even in the presence of some VFI25. Additionally, this study involved professional drivers, whose driving skills are likely superior to those of nonprofessionals, which may have further mitigated the impact of CCDD on MVAs and masked any stronger association.

We observed a significant association between VF-related near misses and MVAs and, after adjusting for potential confounders, the combination of CCDD and VF-related near misses was also significantly associated with MVAs. Tatham et al. demonstrated that even mild VFI can significantly slow reaction times in divided attention tasks during simulated driving, suggesting that early-stage glaucoma may already impair activities such as driving26. These findings suggest that drivers with both CCDD, as glaucoma screening, and a history of VF-related near misses may be at higher risk of MVAs than those without these factors. This combination of CLOCK CHART screening and near-miss history assessment could be valuable in identifying drivers at risk of VFI-related MVA.

In Study 2, of the 326 participants from Study 1 who had CCDD, 14 of the 71 participants (19.7%) who visited an ophthalmologist received a diagnosis of glaucoma. Extrapolating from this proportion, it is possible that approximately 64 of the 326 participants with CCDD could have been diagnosed with glaucoma had they all visited an ophthalmologist. Based on this assumption, the estimated prevalence of glaucoma in this total study population of 1,921 participants is 3.3%. A previous population-based epidemiological study (the Tajimi Study) reported a glaucoma prevalence of 5.0% among Japanese adults aged 40 years and older8. In comparison, the estimated prevalence of 3.3%, derived from a screening program that included younger participants (aged 20–74 years) in an occupational health setting, is lower. This difference is expected, given the demographic differences and the opportunistic nature of screening in this study.

In previous study, the reported percentage of patients with undetected glaucoma varied depending on the populations, ranging from 33 to 98% of primary open-angle glaucoma27,28,29,30,31,32,33,34,35,36,37. The high prevalence of undetected glaucoma is likely due to the lack of subjective symptoms associated with the disease. A recent survey of 227 glaucoma patients, including those with advanced-stage glaucoma, revealed that 65% reported no subjective symptoms while driving6. Glaucoma causes irreversible damage to the VF, underscoring the importance of early diagnosis and intervention to prevent VFI. This study is the first to screen taxi drivers for glaucoma using the CLOCK CHART, suggesting that this VF self-check test could serve as a valuable strategy for detecting asymptomatic glaucoma. This approach has the potential to prevent or reduce motor vehicle accidents associated with VFI.

This study’s major strengths include demonstrating the association between the combination of CCDD and VF-related near misses with MVAs after adjusting for potential confounders and the relatively large sample size of Japanese taxi drivers. However, this study also had several limitations. First, because of its cross-sectional design, we could not establish causal relationships. Further longitudinal studies are necessary to validate these conclusions. Second, most of this data, including those regarding MVAs, were collected through self-administered and onymous questionnaires, potentially introducing reporting and recall biases. Further research incorporating more objective measurement methods for these variables is therefore warranted. Although underestimation of MVAs is often a concern, the proportion of MVAs in this study was higher than in previous studies, suggesting that MVA experiences may not have been underestimated in this study. Third, CLOCK CHART assessments were conducted by multiple examiners, introducing the possibility of inter-examiner bias. However, we mitigated this by training all examiners before the study according to a protocol from a previous study17. Fourth, these results may have been influenced by residual confounding factors such as work conditions, MVA severity, and driving conditions, such as time of day, driver fatigue, traffic conditions, and weather. Finally, this study was conducted within a taxi company, and the proportion of participants who visited an ophthalmologist was relatively low (32.9%) among responders in Study 2. This may indicate potential selection bias, and therefore the representativeness of this study population cannot be guaranteed and generalizing these findings to other populations should be made with caution. Therefore, further longitudinal studies are necessary to adequately address these complexities.

Conclusion

In Study 1, after adjusting for potential confounders, we observed a higher prevalence of MVAs among taxi drivers who had both CCDD and a history of VF-related near miss incidents than in those who did not. However, longitudinal studies are required to determine the causal relationship between VFI and MVAs.

In Study 2, we observed that the CLOCK CHART may aid in identifying undiagnosed glaucoma patients, which may help to prevent MVAs caused by VFI in the future. Further studies using VF self-check tests are warranted to establish a new strategy for early detection of asymptomatic glaucoma.